The present invention relates generally to read-ahead buffers, and more particularly to read-ahead buffers that have been partitioned.
In typical computing system architectures, a server computing device sends messages to a client computing device, at which the messages are consumed by consuming applications. One approach by which messages are sent from the server computing device to the client computing device is for the latter to request messages on a message-by-message basis from the former. In another approach, a read-ahead buffer is employed at the client computing device. Messages are periodically sent, typically in bunches, from the server computing device to the client computing device to fill the read-ahead buffer. The consuming applications then consume the messages from the read-ahead buffer. This latter approach is different than the former approach in that messages are sent from the server computing device to the client computing device in bunches, as opposed to individually on a message-by-message basis as in the former approach.